Rochester Institute of TechnologyRIT Scholar Works

Theses Thesis/Dissertation Collections

1986

Storage of Information on Color PhotographicMaterials with Applications to Optical MemoryDisksAndrew Juenger

Follow this and additional works at: http://scholarworks.rit.edu/theses

This Thesis is brought to you for free and open access by the Thesis/Dissertation Collections at RIT Scholar Works. It has been accepted for inclusionin Theses by an authorized administrator of RIT Scholar Works. For more information, please contact ritscholarworks@rit.edu.

Recommended CitationJuenger, Andrew, "Storage of Information on Color Photographic Materials with Applications to Optical Memory Disks" (1986).Thesis. Rochester Institute of Technology. Accessed from

brought to you by COREView metadata, citation and similar papers at core.ac.uk

provided by RIT Scholar Works

Storage of Information on Color Photographic Materials

with Applications to Optical Memory Disks

Andrew Juenger

April 24, 1986

Revised May 14, 1986

storage of Information on Color Photographic Materials

with Applications to Optical Memory Disks

by

Andrew Juenger

A thesis submitted in partial fulfillment of the requirements for the degree of Bachelor of Science in the Center for

Imaging Science in the College of Graphic Arts and Photography of the Rochester Institute of Technology

Signature of the Author ................................... . Imaging and Photographic Science

Certified Jerald T. LeBlanc by ..........•.........•...•...................... Thesis Advisor

Accepted Ronald Francis by ...........................................••... Supervisor, Undergraduate Research

ii

ROCHESTER INSTITUTE OF TECHNOLOGY

COLLEGE OF GRAPHIC ARTS AND PHOTOGRAPHY

PERMISSION FORM

Title of Thesis: Storage of Information on Color

Photographic Materials with Applications to Optical Memory

Disks

I, Andrew Juenger, hereby grant permission to Wallace

Memorial Library, of R.I.T., to reproduce my thesis in whole

or in part. Any reproduction will not be for commercial use

or profit.

Date: May 14, 1986

111

Storage of Information on Color Photographic Materials

with Applications to Optical Memory Disks

byAndrew Juenger

Submitted to the

Photographic Science and Instrumentation Division

in partial fulfillment of the requirements

for the Bachelor of Science degree

at the Rochester Institute of Technology

ABSTRACT

An experiment was performed to investigate the

information capacity of Ilford Cibachrome Micrographic Film

type CMM.F7, Kodak Aerial Color Film type SO-242, and Kodak

Vericolor Print Film type 4111. The results indicated that

the filmSj-had binary single layer information capacities of

9.51 x 10 , 3.27 x 10 , and 0.87 x 10 bits per square

centimeter respectively. It was estimated that the CMM.F7

material was capable of resolving three density levels for c

trilayer multilevel capacity of 45.2 x 10 bits per square

centimeter, followed by SQ-242 and 4111 with five and six

levels yielding 22.8 x 10 and 6.72 x 10 bits per square

centimeter respectively for trilayer multilevel recording.

Some models for determination of information capacity were

discussed as was application of the experiment to optical

disk recording.

iv

Acknowledgements

The author expresses his sincere thanks to Dr. Jerald

T. LeBlanc of Kodak Research Labs for his help and good

advice on many topics over the past year. Thanks are also

in order for the people in the Materials Coating and

Engineering Division of the Research Laboratory of the

Eastman Kodak Company who helped get the materials and

measurements for this project. Likewise Microcolor

International is thanked for their donation of material.

Finally, a sincere thanks to Maureen, Janet, Bob, and

all the other members of Edge City Productions who have

helped all along.

Table of Contents

I. Introduction 1

A. General 1

B. Spectral Storage 5

C. Information Capacity 8

D. Optical Disk Constraints 10

E. Objective 12

II. Experimental 13

III. Results 18

A. Approach 18

B. Calculated Quantities 26

IV. Discussion 30

V. Conclusions 35

VI. References 37

VII. Appendices 40

VIII. Vita 42

vi

List of Tables

Table 1, Cell area based on line spread function at

0.10 from transformed modulation transfer

function 27

Table 2, Granularity values near 1.0 density measured

with 48 micron diameter circular aperture. . . 27

Table 3, Maximum number of density levels based on G

values and cell areas calculated for the

limiting layers 27

Table 4, Information capacities based on transformed

modulation transfer functions 28

Table 5, Minimum cell area based on line spread function

at 0.10 from edge exposures 28

Table 6, Information capacities from edge exposure. . . 28

Table 7, Minimum cell area based on granularity values. 29

Table 8, Information capacities based on granularities. 29

Table 9, Information capacities of experimental

materials for binary single layer recording. 30

Table 10, Information capacities of several materials

for binary single layer recording 35

vn

List of Figures

Figure 1. Cell size for constant angular velocity. . . 11

Figure 2. Line spread function of Ilford Cibachrome

Micrographic Film type CMM.F7 from

Fourier transform of modulation transfer

function 20

Figure 3. Line spread function of Kodak Aerial

Color Film type SO-242 from Fourier

transform of modulation transfer

function. 21

Figure 4. Line spread function of Kodak Vericolor

Print Film type 41 1 1 from Fourier transform

of modulation transfer function 21

Figure 5. Visual line spread function of Ilford

Cibachrome Micrographic Film type CMM.F7

from edge exposure 22

Figure 6. Visual line spread function of Kodak Aerial

Color Film type SO-242 from edge exposure. 23

Figure 7. Visual line spread function of Kodak Vericolor

Print Film type 4111 from edge exposure. . . 23

Figure 8. One-dimensional illustration of signals

recorded in 1 0 cells with 5 storage densitylevels 25

Vlll

List of Symbols

MTF Modulation transfer function

E(x) Edge Spread function

l(x) Line spread function

FFT Fast Fourier transform

ix

I. Introduction

A. General

With the advent of computing equipment came the

necessity of peripheral systems to provide machine usable

devices for storage of data and operation instructions.

Since the computing systems presently in use operate on

electricity, the presence or absence of charge constitutes

the fundamental delineation of a message. This message unit

is known as a"bit,"

which comes from binary integer. All

information used or stored by a computer system is coded

into binary form. A sequence of bits is used to store

information which is to be accessed by electronic computing

systems for operation or examination by a user. When

entered into a system, or displayed by a system, there is an

intermediate step of translating a symbol from numeric,

character, or pixel format into binary format.

Electronic computer systems facilitate rapid access to

documents stored in peripheral systems. Much information

previously stored on paper in filing cabinets and libraries

is being transferred to machine readable format. Much

information that is generated or collected presently is

going directly into computer operated storage. Together,

the new data and the translation of old documents represent

a tremendous number of bits that must be stored.

The binary representation of information has been

stored on many different media since the first Hollerith

punched card. Punched cards and punched tape were followed

by magnetic storage media in the form of tape, drum, and

disks of various formats. Recently, systems of information

storage by optical media have been developed which are

2beginning to challenge the dominance of magnetic storage.

Among the reasons for the tentative success of the

optical systems are the marketing and operational advantages

of: lower cost of the recording media, greater bit packing

density, and ease of mass duplication. Optical storage

systems have the disadvantage of slower access time and

4transfer rate, and most are not erasable.

One present format of machine readable, optical data

5storage is the disk. First patented in 1965, it was in

1972 that a useful system was first demonstrated. The

basic operation of an optical disk system is the serial

reading of the presence or absence of reflective or

absorptive spots recorded on a spiral track of a rotating

disk. The main system components are: driver, automatic

7focusing optics, reading source, and detector. The spiral

track is analogous to the spiral groove of a phonograph

record. The automatic focuser follows the spiral optical

track.

The mode of spot imaging that is most frequently

described in the literature is that of hole formation by

ablation of thin metalfilms.8' 9' 10

In this system, a

thin layer of a metal that has a low melting point and high

reflectivity is coated onto a substrate, and holes are

burned into the surface with a laser. The burned spot will

not reflect a reading source, while an unburned spot will

reflect. This constitutes the bit. The metal most often

used is tellurium, or alloys of tellurium, and the substrate

is either glass or a polymer such as polymethyl

methacrylate, or polyvinyl chloride. The ablative thin film

method has the advantage of "direct read afterwrite"

capability, which enables a check of the accuracy of

recording within microseconds of the ablation.

A system that is the subject of much research presently

is the phase shift system. This system offers the distinct

advantage of being one of the only erasable optical storage

media. In the phase change media, a film of a material such

as a Te-As-Ge chalcogenide is changed from an amorphous

state to a crystalline state upon exposure to a threshold

level of radiation. When read by a coherent source, the

change in refractive index from amorphous to crystalline

state causes a change in reflectivity and thus defines a

bit. Erasure is done by irradiating the disk at a different

1 1irradiance level. Some other systems include variable

1 o *

birefringence in ferroelectric ceramics and thermomagnetic

1 3development of magnetooptic images.

It is interesting to note that in optical disk

technology, photographic systems are considered

non-conventional. Some photographic systems that have been

investigated are: photoresists, photopolymers, diazo

13 14materials, ultraviolet sensitive tape,

electrophotographic film, and, finally, silver halide

materials.

New silver halide diffusion transfer optical disk

imaging materials have been developed by Drexler, where a

layer of sub-micron diameter silver metal particles are

suspended in a colloid, which is coated over a layer of

silver halide and developing agent molecules in a colloid.

Upon exposure, the silver metal heats up, and melts the

colloid, thus allowing migration to the under layer, where

the metal acts as a catalytic site for nucleation, and

development of the silver halide.' ' '

Fuji Photo Film Co., Ltd. has also developed some

materials expressly for optical disk recording. One such

material combines silver halide with thin metal film,

20 ?1another combines silver halide with dye formation. '

Existing silver halide technology has also been used

effectively. Dry processes such as 3M Dry Silver and Kodak

kind 4117 have been investigated as media for optical data

storage, as have wet process materials such as Kodak film

types 649F, 649GF, and Agfa-Gevaert film types 8E75, 8E56,

10E75, and10E56.13

The tremendous resolution capabilities

of the silver halide wet processes have enabled bit packing

densities rivaled by few other storage media. The obvious

disadvantage is the requirement for image amplification by

development.

All of the systems described are of value to a

particular storage requirement. For purposes of archival

storage of large volumes of unchanging data, the silver

halide system is particularly suitable. The disadvantage of

the need for processing is offset by the high bit density

capability, the long image life, the ability to produce

contact reproductions, and the availability of existing

imaging material technology.

B. Spectral Storage

The use of existing technology can be taken one step

further by introducing spectral storage of data. Consider a

system of dye forming silver halide material such as

transparency material. The spectral absorption of the three

subtractive primary dye layers are used in conjunction to

produce a gamut of reproducible colors.

A system of coding is possible whereby three light

sources of red, green, and blue are used to store data in

the form of colored spots on positive working material. The

simplest case would be the analogy to the present methods,

where irradiance from all three colored sources add to make

a white (clear) spot on the film; no irradiance would

produce a black spot. This is the case of standard binary

representation. However, there are many more combinations

possible. The next consideration is the eight ways that

equal responses of each of the three layers can be utilized

in conjunction with the two remaining layers. Red

illumination alone produces the formation of magenta and

yellow dye, thus forming a red spot. Green and blue

illumination produces a cyan spot. With this method, eight

possibilities can be recorded in any one pixel; they are:

red, green, blue, cyan, magenta, yellow, white, and black.

These eight possibilities represent three bits stored in one

pixel. The response of each layer does not have to be

limited to a pair of absolute values as above; the three

layers could be used in conjunction to produce a set of

possible dominant wavelengths limited only by the

reproducibility of the spectral bandwidth.

The general case of describing the number of possible

ways that a set of signals taken in a certain manner is

shown by raising the number of signal levels to the power of

the number of channels in the signal element. In the

simplest case, described previously as the presence or

absence of a reflective or transmissive spot on a medium,

there is only one channel and there are two levels. The

channel is reflection or transmission, and the levels are

reflecting/transmitting or not reflecting/transmitting. The

number of possible ways to arrange the element is two levels

to the power one channel which is equal to two.

A more illustrative example of this description is

shown by the second case mentioned which is the use of a

trilayer color photographic material. In this case there

are three channels; the cyan, magenta, and yellow dye layers

which are produced imagewise during processing. Since these

are stacked, and interfere with each other only to a minimal

extent, they can be considered as independent channels in

the space of one element. In the simplest case of dye or no

dye present after exposure and processing, there are once

again two levels. The number of possibilities of any one

element is two levels to the power three channels which is

eight, as mentioned earlier.

These eight colors are the equivalent of three black

and white binary bits taken together, and so are equal to

the three dye layers acting independently. If an

intermediate level is introduced corresponding to a mid

density then there are:

-, , -, 3 channels ~- . , . . . . .

3 levels = 27 possibilities

which is equal to the amount of data that three black and

white elements taken together can hold when each has three

levels.

This demonstrates that trilayer material can hold only

three times as much data in the same space, whether taken in

conjunction or independently. It follows then that a

writing and reading scheme can be set up for either mode.

Reading such a pixel to determine the dominant

wavelength could be accomplished by projecting an image of

the information cell onto a diffraction grating focused on

an array of charge couple devices or other photodetectors.

The detector that produces the greatest response indicates

the dominant wavelength once normalization for spectral

response is determined. The independent case is

understandably much simpler. In this mode a record can be

written to one layer by using white light filtered to expose

only that layer. Similarly the remaining two layers can be

exposed. After processing, the three records can be read by

filtered light or three lasers correspondingly.

C. Information Capacity

The desirability of optical storage of data is based

on, among other considerations, its high information

capacity. Information capacity is the measure of how much

data can be stored per unit area of the storage medium. The

units are normally expressed as bits per square centimeter.

The reciprocal of this unit is the area required to store

one bit of information. Thus, the objective of using

photographic material for optical data storage is the

minimization of the size of the cell required to store a bit

of data. This corresponds to the maximization of

information capacity.

Much work has been done by leading scientists to derive

models that can be used to estimate the information capacity

of photographic materials. The impetus for this work is

found in the development of information theory for

22 23electrical signals by Shannon. The work of Jones,

'

McCamy, Altman & Zweig, Shaw, and Lehmbeck will be

discussed more thoroughly in a later section. Each of these

authors developed different but similar models based on

various signal carrying characteristics of photographic

materials. Primary among those functions integral to the

models were the photographic manifestations of signal and

noise characterization given by the spatial functions of

line spread and granularity, and their frequency domain

counterparts of modulation transfer function and Weiner

spectrum.

Generally, the modulation transfer ability of a

photographic material diminishes as spatial frequency

increases. Eventually a limit is approached where the

signal element cannot be distinguished from the noise. The

cell diameter must be large enough to be of a period whose

frequency can be resolved in the modulation transfer

10

function.

Altman and Zweig also gave a representation of the

number of levels that can be used for a given photographic

material based on the available density range and the

variance of the density or granularity. They concluded

that, in most circumstances, binary, or two level, recording

gives the highest information packing density, although in

the use of some materials, for a given acceptable error

rate, many levels can be utilized. This consideration is

discussed in a later section.

D. Optical Disk Constraints

In the context of optical disk storage format, the

information capacity of a material is not necessarily equal

to the reciprocal of the calculated area of a storage cell.

This is due to a particular constraint of the storage

format. Due to obvious engineering considerations, it is

much easier and consistent to have the driving motor

spinning at constant speed, rather than at variable speed.

This implies a constant angular velocity, and as such

requires that the linear velocity is a function of the

radial distance from the center, for any spot on the disk.

gOptical disks normally spin at 1800 rpm. So, a spot that

is twice as far from the center of the disk as another, is

moving twice as fast as the inner spot, since the

circumference is r times greater and both complete a

11

revolution in the same time. This necessitates that the

marks get longer as the radius increases. Therefore, only

the same number of cells can occupy any circumference. This

number is dictated by the number that can fit on the inside

track. Figure 1 illustrates this.

Figure 1 . Cell size for constant angular velocity.

Because of this limitation, the information stored on a

disk is a function of the inside radius, the delta radius

usable, and the minimum cell size. The total number of

elements that can be recorded on a disk is the number that

can be recorded on the inside track times the number of

tracks that can fit within the inner and outer radii. If

the minimum cell diameter, or side, is taken as the period

12

of a complete cycle, then the reciprocal of the side is the

spatial frequency of the cell. It follows, then, that the

number of tracks is the spatial frequency times the delta

radius, and the number of cells per track is the inner

circumference times the spatial frequency- Therefore, the

total number of information elements on a disk is:

# elements = # tracks x # elements per track

= f Ar x f 2 Pi ro

= 2 PirQ

f2

Ar (1 )

The only way to utilize the full information capacity

of the material itself is to use constant linear velocity,

which requires adjusting the speed of the motor as a

function of the position of the reading objective relative

to the center. This introduces much engineering

complication.

E. Objective

The objectives of this experiment were: to determine

the area of the smallest cell that can be made for storage

of information on each of the materials; to determine the

number of density levels that can be utilized for each

material; to calculate the information capacity of these

materials; and to apply the results to the optical disk

model.

13

II. Experimental

Preliminary investigation of this project intuitively

indicated the need for color photographic materials with

high modulation transfer capabilities at high spatial

frequencies. Investigation of published product parameters

and characteristics indicated that the most suitable

materials that were available included: Kodak film type

Aerial Color Film SO-242, Ilford film type Cibachrome

Micrographic Film CMM.F4, and Kodak film type Vericolor

Print Film 41 1 1 .

The first product used was an Ektachrome type reversal

material with incorporated color couplers, the second was a

direct positive dye bleach type material, and the last was a

color negative material. Samples of each were graciously

donated by the manufacturers and distributors.

The initial phase of experimentation was the

development of the necessary processing schemes. Processing

for the color negative material was easily accomplished by

Kodak Process C-41 , likewise the dye bleach material was

processed easily in Cibachrome P-5 chemistry. However, the

Aerial Color Film required the development of a modified

process that included substitutions to Kodak Process EA-5.

This modification was required due to problems with

availability of processing solutions and solution formulae.

A listing of the processing sequence appears in Appendix 1 .

14

In the course of preliminary experimentation it was

discovered that the Cibachrome material showed rapid latent

image fading. Subsequently, all samples were processed

within hours of exposure.

The sensitometric parameters of the materials were then

investigated. While published data of these materials were

available, it was considered important to ascertain the

values of parameters for the experimental conditions. This

was particularly important for the modified process.

Sensitometric parameters were determined from samples

exposed on a Kodak model 1B sensitometer and processed

according to the schemes mentioned above.

The granularity as a function of density was measured

at Kodak Research Labs with a Photometric Data Systems

microdensitometer Model 1000 equipped with 48 micron

diameter aperture at 400x magnification. The values of

sigma density were the mean of the ten lowest values out of

sixteen separate averages of 1000 readings each. The six

highest sample averages were discounted to eliminate error

due to physical defects in the samples. The granularity

values were obtained from samples exposed with a special

step tablet that has minimum inherent grain, which is then

further reduced by diffuse illumination and defocusing.

The modulation transfer functions of the materials were

determined by the use of an instrument designed by R.

15

Lamberts of Kodak ResearchLabs.27

The samples were exposed

directly to sinusoidal radiation distributions of increasing

spatial frequency. The distributions were produced by a

diffraction grating. The exposure modulation was 60% for

all frequencies. Reading the samples was done by the same

instrument operating in the reading mode where the maximum

and minimum density values in the sinusoidal imagewise

density distributions determined the output modulation. The

modulation transfer factor was calculated for each frequency

from the ratio of output modulation to input modulation

thereby determining the modulation transfer function.

An attempt to determine the Weiner spectrum of samples

of each material was unsuccessful due to the fact that each

material was supported on a polyester base. The instrument

that was used before discovery of the problem calculates the

autocorrelation function by optical shifting of an image of

the material microstructure. The optical shifting requires

polarized illumination which was impossible since the

polyester base materials depolarized the light. Had the

autocorrelation functions been determined, the Weiner

spectra would have been calculated by Hankel transformation

of the autocorrelation functions.

Exposures of the edge of a razor blade polished with

mildly abrasive toothpaste were made on each material with a

previous flash exposure such that the density gradient from

16

the high density to the low density portion of the

distribution was within the near linear region of the

characteristic exposure curve for each material. The blades

were tipped on an angle and pressed against the film in

order to maximize contact and produce a microscopically

sharp edge. The processed samples were measured for density

with respect to distance with a PDS microdensitometer

equipped with a 2 x 200 micron rectangular aperture. The

measurements were made at one micron intervals across the

gradient. This served as an approximation to the edge

spread function. The edge spread function was

differentiated to estimate the line spread function.

The line spread function was also estimated by use of

a fast Fourier transform program operating in the back

transform mode. The input data was the modulation transfer

function data from the measurements described previously.

Some experimentation was done on the spectral

sensitivity of the materials by use of a Horton

Spectrosensitometer. The processed samples were compared to

the published data for confirmation.

Preliminary investigation of the use of these materials

in optical disk format included an investigation into

materials for lamination of the photosensitive materials

onto polished aluminum optical disk substrates. Tests were

made with double sided adhesive films to determine their

17

adhesion through the processing solutions. Sima and 3M

brand photo mounting adhesive sheets were tested through all

processes, as was Duro brand five minute epoxy. This

experimentation included development of mounting techniques

for total darkness.

Several visits to the optical disk laboratory of M.

Gupta at Kodak Research Labs served to provide insight into

the considerations for writing to disks laminated with

photographic materials with the equipment available. With

some modifications, the optical disk test bed was used to

write to disk substrates laminated with samples of the

photographic materials. This experimentation was at best

only marginally successful due to difficulties in focusing,

and maintaining focus in total darkness without exposing the

panchromatically sensitive samples.

18

III. Results

A. Approach

As previously mentioned, several models for

determination of the information capacity of photographic

22materials are in the literature. Jones

"

described a

rigorous model based on electrical communication theory

where the calculation of information capacity is given as

one half the integration over frequency space of the base

two logarithm of the quantity one plus the signal to noise

ratio as a function of frequency. This ratio was given as

the Weiner spectrum of the density distribution over the

Weiner spectrum of the granularity distribution.

C = 1/2 log. 1 +

Wd(u,v)

Wg(u,v)

dudv (2)

The model was further developed to include

consideration of the available density range, the gamma of

the material, and the exposure distribution. The

calculation was done in Fourier space to eliminate error due

to "crosstalk"

of the overlap of the spread functions of

adjacent information cells. Since the Weiner spectrum of

the noise due to granularity in the materials was not

determined in the course of this experiment, and since no

analysis of the exposure distribution was made, this

19

particular model was not used. It was mentioned to

illustrate that much rigorous attention has been paid to the

case of information capacity of black and white photographic

materials, and that calculation of information capacity can

be done in the frequency domain as well as the spatial

domain.

The work of Altman and Zweig proved to be the most

useful for analysis of the data from this experiment. The

model described the minimum size of the information cell,

and the maximum number of density levels available based on

some very pragmatic considerations.

The basis of the model was that the minimum cell

dimension must be on the order of the optical spread

function of the imaging material. If the cells were too

small, and packed too close together, the spread function

would cause a contribution of density from one cell to

adjacent cells. Since the spread function of most materials

trails off considerably in the low power region, the edge of

the minimum cell was given as that distance where the spread

function is at a value of 0.10. The diameter of a circular

cell, or the side of a square cell, is then twice that

distance. In this case, the minimum transmittance would be

at most 20% of the maximum transmittance or reflectance in

the event that two"on"

cells were next to each other.

20

Figures 2, 3, and 4 show the estimation of the line

spread functions [l(x)] of each material for each colored

layer from the inverse Fourier transform of the measured

modulation transfer functions.

us -no -see

X RfflD[3@IMi

Figure 2. Line spread function of Ilford Cibachrome

Micrographic Film type CMM.F7 from Fourier transform of

modulation transfer function.

21

qmxdak &jm<iM, @y>H mM s@=s2

Figure 3. Line spread function of Kodak Aerial Color

Film type SO-242 from Fourier transform of modulation

transfer function.

K@AK VHl3ll@lk.@B ^KDIKnT IFIIU_ 4H"D*B

Figure 4. Line spread function of Kodak Vericolor

Print Film type 41 1 1 from Fourier transform of modulation

transfer function.

22

These line spread functions are considered the primary

data for calculation of the information capacity of the

materials investigated. The results are consistent with

those ofLehmbeck26

who used a similar approach to determine

the spread function of an extended range tri- layer color

film.

The results of the differentiation of the edge visual

density profiles are shown below. These spread functions

are given as white light functions, and are considered to be

less reliable than the transformed MTFs for experimental

reasons to be discussed later.

yftfld ^miAE) (F[U)M7D@B!flg _^j_aIF _([%))

0IU>-'IJaiQ> <DAIH](i3@il _0ra@KAl?)IKIIl [MULfifl __U?7

J E

Figure 5. Visual line spread function of Ilford

Cibachrome Micrographic Film type CMM.F7 from edge exposure,

23

BJffll IPBH^I-* tF(U)M7D@Mg _/_a IF E

K[B)AK A1H0AL @IL@K IFDUfl =42

Figure 6. Visual line spread function of Kodak Aerial

Color Film type SO-242 from edge exposure.

is _fl_a if mm

as -m

Figure 7. Visual line spread function of Kodak

Vericolor Print Film type 4111 from edge exposure.

24

The next consideration in the model is the number of

24available density levels for recording. Altman and Zweig

reasoned that since the density at any point on the D Log H

curve is a random variable since it has associated with it

noise due to granularity, the separation in density of the

levels can be chosen such that the probability of reading a

cell as being from one level when it is actually from

another can be calculated from the standard deviation of the

density. A value of + 5 sigma was chosen to insure a bit

6error rate of approximately 10 . This was then doubled as

a safety factor since the distribution of density in a

material is not always normal. Thus density levels must be

separated by 20 times the standard deviation of the density

at a certain level. In this experiment that density was

chosen to be 1.0. Given a certain density range, as well as

a standard deviation, and a multiplier, the number of levels

can be calculated as the range divided by the product of the

standard deviation times twice the multiplier plus one.

Figure 8 illustrates this concept. It is understood that

the standard deviation of density generally increases with

density. Therefore level separation is easier at low

densities than at high densities.

M = (R/2 K sigmaD) +1 (3)

25

It must be noted that M must be an integer. Therefore

all values were rounded to the next lowest integer.

r

Cells

Figure 8. One-dimensional illustration of signals

recorded in 10 cells with 5 storage densitylevels.3

The final method of calculating information capacity

28that was examined was from Dainty and Shaw

,based on

24Altman and Zweig. This model was based on the Selwyn

granularity constant G, given as the density variance at a

particular density times the area of the aperture used to

read the granularity. The density used in this experiment

was at or near 1.0, and the area of the aperture was 1809.6

square microns. Equation (4), below, gives the area of the

smallest cell that can be resolved out of the noise.

J. C. Dainty and R. Shaw, Image Science, Academic

Press, London, 1974, p. 359.

26

A =(2K/R)2

x G (4)

The information capacity is the reciprocal of the area from

equation (4).

B. Calculated Quantities

Tables 1, 2, 3 and 4 below give the calculated

quantities used to determine the information capacities of

the three materials used. These were based on the

estimations of the line spread functions at 0.10 from

transformation of the measured modulation transfer

functions. Density variance is based on the granularity

values measured, and the cell areas calculated. This

implies the assumption that Selwyn's law of granularity

describes the density variance for the cell areas

calculated. Selwyn's law states that the product of

aperture area times density variance is constant over a

range of aperture areas.

2G =

sigmanx Area (5)

From equation (5), the density variances for the

calculated cell areas were calculated from the G values

determined from microdensitometric measurements made with

the 48 micron diameter circular aperture. All calculations

were based on the constraints of the limiting layers for

both cell area and number of levels.

27

Table

modulation

1, Cell areas based on l(x)

transfer functions..10

from transformed

A- (Kx)>10)

Red

Green

Blue

Kx)

CMM.F7

.10

[mics]

10.25

10.25

8.00

[mics ]

105.1

105.1

64.0

SO-242

Kx).10

[mics]

17.50

12.50

11 .50

[mics ]

306.3

156.3

132.3

4111

Kx).10

[mics]

34.0

20.5

13.0

[mics ]

1156

420.3

169.0

Table 2, Granularity values near 1.0 density measured

with 48 micron diameter circular aperture.

G =

sigmaDx A

CMM.F7

Red

G

0.192

Green 0.159

Blue 0.144

SO 242

G

0.139

0.154

0.278

4111

G

0.203

0.312

1.24

Table 3, Maximum number of density levels based on G

values and cell areas calculated for the limiting layers.

M = (R/2K sigmaD) + 1

(R=2.0) CMM.F7

siqma x 1000 M

Red 42.7 3

Green 38.8 3

Blue 37.0 3

(R=3.0) SO-242

siqma x 1000

21.3

22.4

30.1

(R=3.6) 4111

M sigma x 1000 M

8 13.2 14

7 16.4 11

5 32.7 6

28

Table 4, Information capacities based on transformed

modulation transfer functions.

C =

A-1

[bits cm-2]

CMM.F7 SO-242 4111

limiting layer Red/Green Red Red

c 5 51 layer binary capacity 9.51x10? 3.27x1 05

0.87x1 053 layer binary capacity 28.5x10 9.80x10 2.60x10

limiting number of levels 3 c5

56

5M level 3 layer capacity 45.2x10 22.8x10 6.72x10

Tables 5 and 6 below give the calculated quantities for

the experimental materials based on the edge exposures.

Table 5, Minimum cell areas based on line spread

functions at 0.10 from edge exposures.

A =

(Kx)10)2

4~iTT

l(x)1Q

microns 38.4 38 24

Amicrons2

1474 1444 576

Table 6, Information capacities from edge exposures.

-1 -2

C = A [bits cm ]

CMM.F7 SO-242 4111

1 layer binary capacity 0.68x105 0.69x1 05 1.74x10^3 layer binary capacity 2.03x10 2.08x10 5.21x10

Tables 7 and 8 below give the cell areas and

information capacities calculated based on the granularity

values. It should be understood that equation (4) gives the

29

minimum cell area required to resolve the cell from among

the noise due to granularity and does not include

consideration of the line spread function.

Table 7, Minimum cell areas based on granularityvalues.

A =(2K/R)2

x G [microns2]

19.2 6.18 6.27

15.9 6.84 9.63

14.4 12.36 38.3

CMM.F7 SO-242 4111

Red

Green

Blue

Table 8, Information capacities based on granularities,

C = A1[bits cm 2]

CMM.F7 SO-242 4111

limiting layer Red Blue Blue

1 layer binary capacity52.1x10b

80.9x10;? 26.1x1 Or

3 layer binary capacity156x10b

243x10b

78.3x10b

30

IV. Discussion

The following Table 9 is a summary of the results of

this experiment for the binary single layer case.

Table 9, Information capacities of experimental

materials for binary single layer recording.

C =A-1

[bits cm-2]

CMM.F7 SO-242 4111

Capacity from MTF 9.51x10;? 3.27x10;? 0.87x10;?

Capacity from edge 0.68x10;? 0.69x10b

1.74x10;?

Capacity from G 52.1x105

80. 9x105

26.1x10b

The values for information capacity of the three

experimental materials calculated by the granularity

constant G are far too high to be realistic. Such high

values of capacity imply cell areas much smaller than the

areas calculated from the line spread estimations. The line

spread functions estimated are quite similar to those found

in the literature'

for similar color films, so the areas

calculated from equation (4) are assumed to be too low.

Equation (4), as mentioned in the previous section,

defines the minimum cell area for binary recording such that

the cell can be resolved among the noise due to granularity.

In calculating information capacity, there is an implicit

assumption that the cell area can be no smaller than the

minimum for resolution among the noise or within the line

31

spread function. The results described above imply that the

line spread function is the limiting factor in determination

of the minimum cell area, and that granularity controls only

the number of levels available. The fact that noise due to

granularity does not control the minimum cell area is

further evidenced by the observation that for all materials,

multilevel recording of at least three levels is possible

for the given value of K=10. Equation (4) is defined for

binary recording only.

The results indicate that the greatest number of levels

available was six for the Vericolor Print Film Type 4111.

This represents an increase of only 2.6 times the capacity

of binary recording. The smallest number of levels was

three for the Cibachrome Micrographic Film Type CMM.F7

representing only 1.6 times binary capacity. This

demonstrates that since multilevel recording increases

capacity by the base two logarithm of the number of

available levels, reduction of cell area is more critical

than multilevel recording in maximization of information

capacity.

As mentioned previously, the pragmatic approach of

24Altman and Zweig which uses the line spread function to

determine the minimum cell size, is considered the most

useful in this context. Those researchers recommended that

binary recording is the most suitable for photographic

32

materials since the gain in information capacity with

multilevel recording is small compared to the operational

complications. It should be noted that spectral recording

on three dye layers representing three bits per cell

provides greater gain in information capacity than does

multilevel recording on the three materials examined.

As described, the cell area was determined by the width

of the 10% value of the line spread function. The line

spread function of each film was calculated by two methods

mentioned previously. The transformation of the modulation

transfer function is considered more accurate because the

experimental procedure for the edge exposures had inherent

error due to reflection off the surface of the razor blades.

While every effort was made to produce accurate results

by polishing the blades with mildly abrasive toothpaste, and

maximizing intimate contact between the blades and the film,

the precaution of using black razor blades was not realized

until the result of increased spread presumably from

reflection off the blade surface was discovered by the

observed asymmetry in the resulting spread functions. It

was the intention of this experiment to use the edge

exposures as a supplemental method for determination of the

line spread function. The transform of the modulation

transfer function was intended to be the primary source of

data for the determination of the information capacity of

33

the experimental materials, and the results indicate that

this was prudent.

There are some artifacts of the transformation program

evident in the line spread functions calculated from the

modulation transfer functions. These are manifested as wavy

trails towards the outer edges of some of the spread

functions and are presumed to be from the requirement of

using discrete approximations to the continuous functions.

In the cases where the wavyness occurred in the region of

0.10 spread which defines the cell width, a smooth curve was

fitted to the data.

A second consideration of the results of the

transformation is the fact that the phase was assumed to be

zero. The optical transfer function which includes both

modulus and phase would have been more precise; however, the

assumption that the phase was unshifted is reasonable since

phase shift normally does not occur until near the first

zero of modulation transfer for symmetric functions such as

the sinusoid used for this experiment.

The results of the experiment in the case of the

transformed data are considered valid. The data results

from the average of four very similar samples of each

material. A first order approximation of the estimated

uncertainty of the results of this experiment based on the

standard deviation of a representative sample of MTF data

34

indicates that for this experiment, the results are precise

to within 18% for 99% confidence. This is a conservative

estimate. The calculated information capacity seems

reasonable in comparison to previously published results of

other experiments.

Disks laminated with these materials were written to

with great operational difficulty. No actual information

was recorded but elements were generated on each of the

three colored layers independently. The difficulty was

primarily due to the fact that the experimental test bed was

designed to write to disks of considerably different design

and sensitivity. As described previously, focusing was the

greatest source of difficulty. Future work would include

the assembly of a different test bed which is a considerably

expensive and precise task.

35

V. Conclusions

The information capacities of the experimental

materials for binary single layer recording calculated from

the modulation transfer function data are shown below in

Table 10 with the results of previously published data for

other materials as well as the information capacity of

Tellurium disk material based on the standard pitch of 1.8

microns as the cell width.

Table 10, Information capacities of several materials

for binary single layer recording.

C = A1[bits cm-2]

Material C

Tellurium Disk308x10"

Ilford CMM.F7 9.51x10;?

KodakPanatomic-Xa

4. 4x10;?

Kodak SO-2423.27x10b

Kodak 4111 . 0.87x10;?

Kodak Extended Range 0.82x10

The information capacity of each of the three films

investigated was limited by the size of the spread function

of the cyan dye layer. The results of the transformation of

the modulation transfer functions was considered most

accurate and produced results that were quite acceptable in

comparison to previous research. The Ilford Micrographic

Altman and Zweig, Photo. Sci. Eng., 6:174 (1963)

bLehmbeck, Photo. Sci. Eng., 11:270 (1967).

36

film had the highest binary information capacity at 9.51 x

10 bits per square centimeter for a single layer which is

24higher than Panatomic-X as reported by Altman and Zweig as

4.4 x 105. Kodak Aerial Color film followed at 3.27 x 105,

Kodak Vericolor print film at 0.87 x 10 bits per square

2 6centimeter was just above the value reported by Lehmbeck

of 0.82 x105

for Kodak Extended Range Film. Direct

comparison with the published data must be considered as

only an indication of relative performance since the

experimental procedures were quite different. The actual

capacity can be somewhat higher with multilevel trilayer

recording as shown in the results. In comparison to

Tellurium disk material with a cell width of 1.8 microns,

and information capacity of 308x10 bits per square

5centimeter, the best trilayer capacity of 28.5x10 bits per

square centimeter for the CMM.F7 represents only 9.3% of the

capacity of present disk materials.

The application of color photographic materials to

optical disk recording may not be practical considering the

disadvantages of the requirement for processing, and the

complications of spectral reading and writing. Nevertheless

the experiment served to confirm that modern color

photographic materials have very high information capacity

that may be utilized in other contexts, including

photomicrography -

37

References

1. C. S. McCamy, "On the Information in a

Microphotograph"

, AppI. Optics, 4, p. 408 (1965).

2. P. Popoff and J. Ledieu, "Towards New Information

Systems: Gigadisc", Applications of Optical Digital

Data Disk Storage Systems, W. M. Deese, M. Carasso,

Eds., Proc. SPIE 490, p. 21 (1984).

3. M. W. Goldberg, "Large Memory Applications for Optical

Disk", Optical Data Storage Technical Digest, Opt. Soc.

Am., p. MA3-1 (1983).

4. L. Fujitani, "Laser Optical Disk: The Coming Revolution

in On-Line Storage", G. Dallaire, Ed., Commun. ACM.,

27, p. 547 (1984). (Editor's Note.)

5. F. F. Dove, U.S. Pat. No. 3,226,696 (1965).

6. J. Isailovic, Videodisc and Optical Memory Systems,

Prentice-Hall, Inc., Englewood Cliffs, N.J., 1985, p.

5.

7. V. B. Jipson and K. Y. Ahn, "Materials for Optical

Storage", Solid State Tech., 27, p. 141 (1984).

8. P. Kivits, B. Jacobs, and P. Zalm, "Summary Abstract:Research on Materials for Optical Storage", Optical

Storage Materials, T. H. DiStefano, Ed., Proc. SPIE

263, p. 68 (1980).

9. J. Corcoran and H. Ferrier, "Melting Holes in Metal

Films for Real-Time, High Density Digital Data

Storage", Optical Storage Materials and Methods, L.

Beiser, D. Chen, Eds., Proc. SPIE 123, p. 17 (1977).

10. R. McFarlane, et al. , "Digital Optical Recorders at 5

Mbit/s Data Rate", Opt. Eng., 21, p. 913 (1982).

11. A. W. Smith, "Injection Laser Writing on Chalcogenide

Films", Optical Storage of Digital Data Technical

Digest, Opt. Soc. Am., p. MB6-1 (1973).

12. N. Bar-Chaim, A. Seidman, and E. Wiener-Avnear, "A

Color Memory Mode Based on the Variable Birefringence

in PLZT Ferroelectric Ceramics", Ferroelectrics , 11,

pp. 385-388 (1976).

38

13. R. A. Bartolini, H. A. Weakliem, and B. F. Williams,"Review and Analysis of Optical Recording Media", Opt.

Eng.. 15, pp. 99-105 (1976).

14. F. N. Magee, "Discussion of Electrophotographic Film",Laser Recording and Information Handling, A. A.

Jamberdino, Ed., Proc. SPIE 200, pp. 16-19 (1979).

15. S. Maslowski, "High Density Data Storage on Ultraviolet

Sensitive Tape", Optical Storage of Digital Data

Technical Digest, Opt. Soc. Am., p. WA4-1 (1973).

16. J. Drexler,"Drexon

Optical Memory Media for Laser

Recording and Archival Data Storage", J. Vac. Sci.

Technol . , 18, pp. 87-91 (1981).

17. J. Drexler and E. W. Bouldin, U.S. Pat. No. 4,312,938

(1982).

18. E. W. Bouldin and J. Drexler, U.S. Pat. No. 4,298,684

(1981 ).

19. J. Drexler and E. W. Bouldin, U.S. Pat. No. 4,284,716

(1981 ).

20. R. G. Zech, "Review of Optical Storage Media", Optical

Information Storage, K. G. Lieb, Ed., Proc. SPIE 177,pp. 56-62 (1979).

21. Y. Mukunoki, et al., U.S. Pat. No. 4,278,797 (1981).

22. R. C. Jones, "information Capacity of Photographic

Films", J. Opt. Soc. Am., 51, pp. 1159-1171 (1961).

23. R. C. Jones, "On the Point and Line Spread Functions of

Photographic Images", J. Opt. Soc. Am., 48, pp. 934-937

(1958).

24. J. H. Altman and H. J. Zweig, "Effect of Spread

Function on the Storage of Information on Photographic

Emulsions", Photo. Sci. Eng., 7, pp. 173-177 (1963).

25. R. Shaw, "The Application of Fourier Techniques and

Information Theory to the Assessment of Photographic

Image Quality", Photo. Sci. Eng., 6, pp. 281-286

(1962).

39

26. D. R. Lehmbeck, "Experimental Study of the Information

Storing Properties of Extended Range Film", Photo. Sci.

Eng., 11, pp. 270-278 (1967).

27. R. L. Lamberts and F. C. Eisen, "A System for Automated

Evaluation of Modulation Transfer Functions of

Photographic Materials", J . AppI . Phot . Eng . , 6, p. 1

(1980).

28. J. C. Dainty and R. Shaw, Image Science, Academic

Press, London, 1974, pp. 344-376.

40

'2

Appendix 1

Modified Process ME-4 at 97.5 F

Step time agitation

1. ME-4* Prehardener 2'30" 2"<ai0"

N.

2. ME-4* Neutralizer30"

3. ME-4* 1st Developer2*06" 2"@10"

N-

4. 1st Stop** 30"

5. Wash 1'00"

6. ME-4* Color Developer6' 00" 2"@10"

N

7. 2nd Stop** 30" 2"@10"

N,8. Wash

1'00"

9. K-12A*** Bleach1'30"

continuous air

10. E-6**** Fix1'30" 2"@10"

N911. Wash

2' 00"

12. e-6**** Stabilizer30"

* Kodak Process ME-4

** 1:21 28% glacial acetic acid

*** Kodak Process K-12A

**** Kodak Process E-6

41

Appendix 2

Emulsion numbers of material samples

Ilford Cibachrome Micrographic Film Type CMM.F7

Emulsion number 54A301 -3942

Kodak Aerial Color Film Type SO-242

Emulsion number 74-752B

Kodak Vericolor Print Film Type 4111

Emulsion number 482 21 4

42

Vita

Andrew Juenger was born in Easton, Pennsylvania in

1961. Having been interested in the creative aspects of

photography for most of his life, Andrew chose to study the

more challenging technical discipline of Imaging Science to

provide a complete and rigorous education in the field of

imaging to complement his education and professional

experience in photography. Andrew has been working since

January 1 985 as a technician for Eastman Kodak Company in

the Materials Coating and Engineering Division of the

Research Laboratory.